Psychotic symptoms occur in approximately 50% of individuals diagnosed with Alzheimer Disease (AD+Psychosis, AD+P). AD+P is associated with greater cognitive decline and increased institutionalization. Current therapies have limited benefit for AD+P, and do not alter its poor prognosis. We have previously estimated the heritability of psychosis in AD as 61%-70%, indicating a substantial genetic component. Recently linkage and/or association of several novel genes with idiopathic psychosis have been reported, two of which (NRG1 and COMT) we have found to be associated with AD+P in preliminary studies. Other data suggest that variation in genes contributing to neurodegenerative pathology itself (e.g. APP, PS1, and MAPI) also alter psychosis risk. We now propose to analyze this highly probable set of candidate genes to address several questions regarding AD+P: 1) Which genes demonstrating linkage and allelic association with idiopathic psychosis, or leading to neurodegenerative pathology, increase risk for AD+P?;2) What are the effects of variation in these genes on predicting psychosis onset during AD, and how do these effects interact with cognitive impairment to increase AD+P risk?;and, 3) ls there evidence for subtypes within AD+P and how are they influenced by genotype? We will address these questions in three aims, involving three cohorts. Single locus and haplotype associations of NRG1, DTNBP1, DISC1, COMT, DAOA (formerly G72), AKT1, RGS4, APP, PS1, PS2, and MAPI with AD+P will be tested in a large (N=1000) AD+P vs AD-P Case-Control Cohort. Significant associations will be confirmed in a similarly large Family Cohort. Confirmed associations will be further evaluated for the prediction of AD+P onset in a Prospective Cohort of 786 AD and Mild Cognitive Impairment subjects without psychosis at study entry. Mediating and moderating interactions between genetic variation and cognition on AD+P onset will be evaluated. Longitudinal data from the Prospective Cohort will be used to identify subtypes of AD+P with differing trajectories and genetic determinants. Successful completion of these aims will identify genetic predictors of AD+P, indicate if they act via an effect on cognition, and evaluate if they affect all individuals uniformly or result in subtypes. These findings may guide the development of interventions for the prediction, treatment, and/or prevention of psychosis and excess cognitive morbidity in AD.